PostgreSQL, DuckDB, Spark, Databricks: Accept a comparison in the BETWEEN lower bound

Author: LucaCappelletti94

src/dialect/databricks.rs

@@ -44,6 +44,10 @@ impl Dialect for DatabricksDialect {
         false
     }
 
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        true
+    }
+
     /// See <https://docs.databricks.com/en/sql/language-manual/sql-ref-identifiers.html>
     fn identifier_quote_style(&self, _identifier: &str) -> Option<char> {
         Some('`')

src/dialect/duckdb.rs

@@ -65,6 +65,10 @@ impl Dialect for DuckDbDialect {
         true
     }
 
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        true
+    }
+
     fn supports_named_fn_args_with_eq_operator(&self) -> bool {
         true
     }

src/dialect/generic.rs

@@ -173,6 +173,10 @@ impl Dialect for GenericDialect {
         true
     }
 
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        true
+    }
+
     fn supports_comment_on(&self) -> bool {
         true
     }

src/dialect/mod.rs

@@ -1089,6 +1089,17 @@ pub trait Dialect: Debug + Any {
         self.prec_value(Precedence::Between)
     }
 
+    /// Returns true if the low bound of a `BETWEEN` predicate may end with one comparison,
+    /// as in `a BETWEEN b > 1 AND c`, while `IS`, `LIKE` and `BETWEEN` still end the bound.
+    ///
+    /// The low bound is a `b_expr` in the
+    /// [PostgreSQL grammar](https://github.com/postgres/postgres/blob/master/src/backend/parser/gram.y)
+    /// and a `valueExpression` in the
+    /// [Spark grammar](https://github.com/apache/spark/blob/master/sql/api/src/main/antlr4/org/apache/spark/sql/catalyst/parser/SqlBaseParser.g4).
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        false
+    }
+
     /// Returns true if this dialect requires the `TABLE` keyword after `DESCRIBE`
     ///
     /// Defaults to false.

src/dialect/postgresql.rs

@@ -412,6 +412,10 @@ impl Dialect for PostgreSqlDialect {
         true
     }
 
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        true
+    }
+
     /// see <https://www.postgresql.org/docs/current/sql-comment.html>
     fn supports_comment_on(&self) -> bool {
         true

src/dialect/spark.rs

@@ -48,6 +48,10 @@ impl Dialect for SparkSqlDialect {
         false
     }
 
+    fn supports_comparison_in_between_low_bound(&self) -> bool {
+        true
+    }
+
     /// See <https://spark.apache.org/docs/latest/sql-ref-identifier.html>
     fn identifier_quote_style(&self, _identifier: &str) -> Option<char> {
         Some('`')

src/parser/mod.rs

@@ -4576,7 +4576,22 @@ impl<'a> Parser<'a> {
 
     /// Parses `BETWEEN <low> AND <high>`, assuming the `BETWEEN` keyword was already consumed.
     pub fn parse_between(&mut self, expr: Expr, negated: bool) -> Result<Expr, ParserError> {
-        let low = self.parse_subexpr(self.dialect.between_low_bound_precedence())?;
+        let mut low = self.parse_subexpr(self.dialect.between_low_bound_precedence())?;
+        if self.dialect.supports_comparison_in_between_low_bound()
+            && matches!(
+                self.peek_token_ref().token,
+                Token::Eq
+                    | Token::DoubleEq
+                    | Token::Neq
+                    | Token::Lt
+                    | Token::LtEq
+                    | Token::Gt
+                    | Token::GtEq
+                    | Token::Spaceship
+            )
+        {
+            low = self.parse_infix(low, self.dialect.prec_value(Precedence::Between))?;
+        }
         self.expect_keyword_is(Keyword::AND)?;
         let high = self.parse_subexpr(self.dialect.prec_value(Precedence::Between))?;
         Ok(Expr::Between {

tests/sqlparser_common.rs

@@ -21022,3 +21022,55 @@ fn parse_window_spec_keyword_base_window_name() {
         "SELECT count(*) OVER (ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) FROM t",
     );
 }
+
+#[test]
+fn parse_comparison_in_between_low_bound() {
+    let dialects = all_dialects_where(|d| d.supports_comparison_in_between_low_bound());
+    for (sql, expected) in [
+        ("a BETWEEN b > 1 AND c", BinaryOperator::Gt),
+        ("a BETWEEN b <> 1 AND c", BinaryOperator::NotEq),
+        ("a BETWEEN b + 1 >= 2 AND c", BinaryOperator::GtEq),
+    ] {
+        match dialects.verified_expr(sql) {
+            Expr::Between { low, .. } => assert!(
+                matches!(&*low, Expr::BinaryOp { op, .. } if *op == expected),
+                "{sql}: {low:?}"
+            ),
+            other => panic!("{sql}: expected BETWEEN, got {other:?}"),
+        }
+    }
+    for (sql, error) in [
+        (
+            "SELECT a BETWEEN b = 1 = true AND c",
+            "Expected: AND, found: =",
+        ),
+        (
+            "SELECT a BETWEEN b LIKE 'x' AND c",
+            "Expected: AND, found: LIKE",
+        ),
+        (
+            "SELECT a BETWEEN b IS NULL AND c",
+            "Expected: AND, found: IS",
+        ),
+        (
+            "SELECT a BETWEEN b BETWEEN 1 AND 2 AND c",
+            "Expected: AND, found: BETWEEN",
+        ),
+    ] {
+        assert_eq!(
+            dialects.parse_sql_statements(sql),
+            Err(ParserError::ParserError(error.to_string())),
+            "{sql}"
+        );
+    }
+
+    assert_eq!(
+        all_dialects_except(|d| {
+            d.supports_comparison_in_between_low_bound() || d.is::<SQLiteDialect>()
+        })
+        .parse_sql_statements("SELECT a BETWEEN b > 1 AND c"),
+        Err(ParserError::ParserError(
+            "Expected: AND, found: >".to_string()
+        ))
+    );
+}

tests/sqlparser_sqlite.rs

@@ -1185,21 +1185,12 @@ fn parse_between_low_bound_precedence() {
     sqlite().verified_expr("5 BETWEEN 1 AND NULL IS NULL");
     sqlite().verified_expr("5 BETWEEN 1 AND 10 = 1");
 
-    // The inner AND belongs to the bound expression, so the missing `)` is reported at FROM.
     assert_eq!(
         sqlite().parse_sql_statements("SELECT a BETWEEN (1 AND 2 FROM t"),
         Err(ParserError::ParserError(
             "Expected: ), found: FROM".to_string()
         ))
     );
-
-    assert_eq!(
-        TestedDialects::new(vec![Box::new(GenericDialect {})])
-            .parse_sql_statements("SELECT a BETWEEN b > 1 AND c FROM t"),
-        Err(ParserError::ParserError(
-            "Expected: AND, found: >".to_string()
-        ))
-    );
 }
 
 fn sqlite() -> TestedDialects {